Recently, systems which use multimedia contents such as moving images, still images, sounds, and documents have been implemented along with the development of digital data processing/communication technologies. Among these systems, there is also proposed a system which can create and record various multimedia contents and uses multimedia contents containing various types of multimedia contents.
As multimedia contents are used more and more, the data size of each multimedia content increases with an increase in expressions such as the resolution of the multimedia content.
When different save formats appropriate for recording and use are selected for respective multimedia contents, various formats coexist. A device which utilizes such multimedia contents must be compatible with many formats, as needed.
As multimedia contents are exploited more and more, the user utilizes and accumulates many multimedia contents, and the search technique for multimedia contents becomes more important.
The search technique must deal with especially many formats of multimedia contents and a large data size. In addition, to cope with a large amount of multimedia contents, the search technique has changed from a technique of simply reading out internal information of each multimedia content and searching for the multimedia content.
As the search technique, it becomes popular to extract information necessary for search as metadata in advance from the contents of each multimedia content in accordance with each moving image, still image, sound, document, or the like contained in the multimedia content, accumulate the metadata, and utilize the metadata for search.
According to this method, if extracted/saved metadata have the same format, the metadata can be searched for regardless of the format of each multimedia content in search. As a characteristic feature of this method, a large amount of multimedia contents can be searched for at a high speed without reading and interpreting internal information of a large amount of multimedia contents at a large data size in search.
As the digital communication technique for a network and the like advances, the environment is ripe for a process which can be performed in cooperation with a plurality of devices. However, even with an increase in communication speed by the development of the communication technique, a long time is taken to transfer multimedia contents of a relatively large data size.
Even in this environment, transfer of multimedia contents puts a heavy load on the device and network. In general, even the above-mentioned system which saves and shares multimedia contents does not freely transfer multimedia contents.
This phenomenon becomes noticeable particularly when only a low-speed network environment can be used or the performance of a user's device is low. In many cases, devices which save and share multimedia contents independently exist and are distributed on the network.
The search technique itself can extract and save, with a common format and items, metadata necessary to search for a single media content such as a still image or sound as far as the metadata is simple such as a creation date and time or a title. However, most of items and formats which depend on media contents cannot be commonly used, and search techniques and usages of their metadata are greatly different from each other.
For example, for a moving or still image, the main color of an image information part of the content or the like can be extracted as a feature amount (quantity) and used as metadata. For sound, the frequency component of an audio information part can be extracted as a feature amount and used as metadata. These metadata have nothing in common, and search functions for these data are implemented in different ways.
For this reason, services for providing the search function and the like are often independently provided as search functions corresponding to respective media. These search functions are distributed on the network.
In the above-described environment, when multimedia contents are searched for, a search service capable of effectively searching them for a still image part and a search service capable of effectively searching them for a text part such as a document content or title are different. These search services are distributed on the network.
Owing to the convenience of the system or network, devices which can possess and share target multimedia contents may be limited depending on the search service. In this environment, the user must search for a search service which can search for desired multimedia contents.
Although standards such as WSDL (Web Services Description Language by the W3C) which targets search for a search service have been established, respective search services have different functions, as described above. Further, there is neither information on multimedia contents which can be searched for by the service, nor information on devices which save and share the multimedia contents. Thus, the above problems cannot be solved.
As a method for solving these problems, several techniques have been proposed.
For example, in Japanese Patent Laid-Open No. 2001-195421, search index information is kept accurate while storages/search engines are distributed. In order to accurately reflect in the search engine a change of content information in a dynamic storage, the update data is exchanged. However, in Japanese Patent Laid-Open No. 2001-195421, the relationship between search engines and their recognition must be given in advance.
In Japanese Patent Laid-Open No. 2001-160068, a search instruction used for search is recorded as a search target by adding metadata such as a comment. By allowing reuse of search instruction information, a search instruction can be easily input.
However, Japanese Patent Laid-Open Nos. 2001-195421 and 2001-160068 do not consider any configuration for using a search engine optimal for a search instruction from a plurality of types of search engines, and cannot execute efficient search.